Apparatuses and systems for photon detection can include a first optical sensing structure structured to absorb light at a first optical wavelength; and a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures. The second optical sensing structure can be structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure. Apparatuses and systems can include a bandgap grading region.
|
25. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength;
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure, and
a dielectric layer interfacing between the first and the second optical sensing structures to permit transmission of light and to fuse the first and the second optical sensing structures together as a single structure.
26. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength; and
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure,
wherein a dead time of the second optical sensing structure overlaps with an avalanche time of the first optical sensing structure to prevent a positive feedback loop between the first and second optical sensing structures.
22. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength; and
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure,
wherein the second optical sensing structure comprises an absorption structure which absorbs light at the second optical wavelength and a multiplication structure between the absorption structure and the first optical sensing structure to emit light at the first optical wavelength.
34. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength; and
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure,
wherein the first optical sensing structure comprises an array of detector pixels, wherein the second optical sensing structure comprises an array of detector pixels, the device further comprising one or more layers to minimize inter-pixel cross-talk between the first and second optical sensing structures.
16. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength; and
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure,
wherein the first optical sensing structure comprises a silicon substrate and is a silicon-based optical detector,
wherein the second optical sensing structure is an ir optical detector which emits light at the first optical wavelength by converting energy in absorbed light at the second optical wavelength via luminescence resulting from hot carrier recombination.
1. A semiconductor radiation sensing device, comprising:
a semiconductor absorption region structured to absorb photons at a first wavelength to generate one or more charged carriers;
a multiplication region structured to receive the one or more charged carriers, the multiplication region structured to generate an avalanche of electrons in response to the one or more charged carriers and emit secondary photons at a second wavelength shorter than the first wavelength;
a buffer region structured to impede electrons or holes from the avalanche from passing through the buffer region to cause a reduction in an electric field across the multiplication region to quench the avalanche; and
a bandgap grading region adjacent to the absorption region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the absorption region and a second region.
27. A semiconductor device, comprising:
a first optical sensing structure structured to absorb light at a first optical wavelength; and
a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures, wherein the second optical sensing structure is structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure,
wherein the second optical sensing structure comprises:
an absorption region structured to absorb photons at the second wavelength,
a multiplication region structured to generate an avalanche of electrons in response to an absorbed photon to cause photons to be emitted at the first optical wavelength, and
a buffer region coupled with the multiplication region, and structured to impede electrons or holes from the avalanche from passing through the buffer region to cause a reduction in an electric field across the multiplication region to quench the avalanche, and to allow electrons or holes to pass through the buffer region to cause an increase in the electric field across the multiplication region to facilitate a recovery from the avalanche.
14. A semiconductor radiation sensing device, comprising:
a semiconductor absorption region structured to receive light at a first wavelength to generate one or more charged carriers by absorbing received light;
a multiplication region structured to receive the one or more charged carriers generated from the semiconductor absorption region and to generate an avalanche of secondary charged carriers in response to the one or more charged carriers and emit secondary photons from the secondary charged carriers;
a region coupled between the absorption region and the multiplication region, the region comprising a mechanism that quenches the avalanche of the multiplication region after occurrence of the avalanche and resets the multiplication region for a next avalanche; and
a semiconductor transition region formed between the buffer region and the absorption region to have a first bandgap at a first interface with the buffer region that is equal to or similar to a bandgap of the buffer region and a second bandgap at a second interface with the absorption region that is equal to or similar to a bandgap of the absorption region, the semiconductor transition region having a spatially varying bandgap between the first and second interfaces to eliminate an abrupt change in bandgap between the buffer region and the absorption region.
12. A semiconductor radiation sensing device, comprising:
a semiconductor absorption region structured to receive light at a first wavelength to generate one or more charged carriers by absorbing received light;
a multiplication region structured to receive the one or more charged carriers generated from the semiconductor absorption region and to generate an avalanche of secondary charged carriers in response to the one or more charged carriers and emit secondary photons from the secondary charged carriers;
a region coupled between the absorption region and the multiplication region, the region comprising a mechanism that quenches the avalanche of the multiplication region after occurrence of the avalanche and resets the multiplication region for a next avalanche; and
a semiconductor transition region formed between the multiplication region and the absorption region to have a first bandgap at a first interface with the multiplication region that is equal to or similar to a bandgap of the multiplication region and a second bandgap at a second interface with the absorption region that is equal to or similar to a bandgap of the absorption region, the semiconductor transition region having a spatially varying bandgap between the first and second interfaces to eliminate an abrupt change in bandgap between the multiplication region and the absorption region.
2. The device as in
3. The device as in
4. The device as in
5. The device as in
6. The device as in
7. The device as in
a mechanism to bias the second optical sensing structure at a DC voltage.
8. The device as in
10. The device as in
11. The device as in
13. The device as in
15. The device as in
17. The device as in
a silicon dielectric layer as an interfacing layer formed between the first optical sensing structure and the second optical sensing structure.
18. The device as in
a complementary metal-oxide-semiconductor (CMOS) circuit to control the first optical sensing structure and the second optical sensing structure.
19. The device as in
20. The device as in
21. The device as in
23. The device as in
24. The device as in
28. The device as in
a bandgap grading region coupled with the multiplication region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the multiplication region and a second region.
29. The device as in
a bandgap grading region coupled with the buffer region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the buffer region and a second region.
30. The device as in
a mechanism to bias the second optical sensing structure at a DC voltage.
31. The device as in
33. The device as in
|
This document claims the benefit of U.S. Provisional Application No. 61/080,172 entitled “Integrated IR Detector” and filed on Jul. 11, 2008, which is incorporated by reference as part of the disclosure of this document.
The invention was made with government support under Grant No. W911NF-05-1-0243 awarded by ARO. The government has certain rights in the invention.
This document relates to semiconductor devices including semiconductor photodetectors and photodetector arrays.
Semiconductor pn junctions can be used to construct photodiodes for detecting photons. An avalanche diode is one example of such photodiodes. Single-photon avalanche diodes (SPADs) are designed to detect single photons and can be used in a variety of applications, including biological, military and biometric applications. A SPAD can be operated in a Geiger mode to output a signal that represents a photon-arrival event or the time of arrival of a photon. As such, SPADs can be used as low-light-level imagers, sensors for Time-Correlated Single-Photon Counting and Fluorescence Correlation Spectroscopy and other sensing applications. When operated in the Geiger mode, a SPAD is in essence a reverse-biased pn junction which can sustain numerous avalanche breakdowns without incurring damage and with minimal charge trapping. When a photon is absorbed in the high-field region of the SPAD, the SPAD generates an electron-hole pair which can induce an avalanche through impact ionizations. This avalanche can be electrically sensed with high timing accuracy, and is quickly quenched. The pn junction is then reactivated by recharging the junction in excess of its breakdown voltage.
This document describes technologies, among other things, for single photon detection.
In one aspect, apparatuses and systems for photon detection can include a first optical sensing structure structured to absorb light at a first optical wavelength; and a second optical sensing structure engaged with the first optical sensing structure to allow optical communication between the first and the second optical sensing structures. The second optical sensing structure can be structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure.
These and other implementations can include one or more of the following features. The first optical sensing structure can include a silicon substrate and is a silicon-based optical detector. The second optical sensing structure can include an IR optical detector which emits light at the first optical wavelength by converting energy in absorbed light at the second optical wavelength via luminescence resulting from hot carrier recombination. The first optical sensing structure can include a silicon detector. The second optical sensing structure can include a III-V semiconductor stack. The second optical sensing structure can include an InGaAs/InP stack.
In some implementations, apparatuses and systems can include a silicon dielectric layer as an interfacing layer formed between the first optical sensing structure and the second optical sensing structure. Apparatuses and systems can include a complementary metal-oxide-semiconductor (CMOS) circuit to control the first optical sensing structure and the second optical sensing structure.
In some implementations, the second optical sensing structure can include an absorption structure which absorbs light at the second optical wavelength and a multiplication material layer between the absorption structure and the first optical sensing structure to emit light at the first optical wavelength. The absorption structure can have a bandgap less than a bandgap of the multiplication structure. The absorption structure can have a bandgap similar to a bandgap of the multiplication structure.
In some implementations, the second optical sensing structure can include an absorption structure which absorbs light at the second optical wavelength and a multiplication structure between the absorption structure and the first optical sensing structure to emit light at the first optical wavelength. The absorption structure can have a bandgap less than a bandgap of the multiplication structure. The absorption structure can have a bandgap similar to a bandgap of the multiplication structure.
In some implementations, apparatuses and systems can include a dielectric layer interfacing between the first and the second optical sensing structures to permit transmission of light and to fuse the first and the second optical sensing structures together as a single structure. The first and the second optical sensing structures can be fused together as a single structure are free of an electrical contact structure between the first and the second optical sensing structures. The first optical sensing structure can include a SiO2 shallow-trench-isolation (STI) guard ring structure.
The first optical sensing structure can include an array of detector pixels. The second optical sensing structure can include an array of detector pixels. Implementations can include one or more layers to minimize inter-pixel cross-talk between the first and second optical sensing structures.
The second optical sensing structure can include an absorption region structured to absorb photons at the second wavelength; a multiplication region structured to generate an avalanche of electrons in response to an absorbed photon to cause photons to be emitted at the first optical wavelength; and a buffer region coupled with the multiplication region, and structured to impede electrons or holes from the avalanche from passing through the buffer region to cause a reduction in an electric field across the multiplication region to quench the avalanche, and to allow electrons or holes to pass through the buffer region to cause an increase in the electric field across the multiplication region to facilitate a recovery from the avalanche. Implementations can include a bandgap grading region coupled with the multiplication region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the multiplication region and a second region.
In yet another aspect, apparatuses and systems can include a semiconductor absorption region structured to receive light at a first wavelength to generate one or more charged carriers by absorbing received light; a multiplication region structured to receive the one or more charged carriers generated from the semiconductor absorption region and to generate an avalanche of secondary charged carriers in response to the one or more charged carriers and emit secondary photons from the secondary charged carriers; a region coupled between the absorption region and the multiplication region, the region comprising a mechanism that quenches the avalanche of the multiplication region after occurrence of the avalanche and resets the multiplication region for a next avalanche; and a semiconductor transition region formed between the multiplication region and the absorption region to have a first bandgap at a first interface with the multiplication region that is equal to or similar to a bandgap of the multiplication region and a second bandgap at a second interface with the absorption region that is equal to or similar to a bandgap of the absorption region, the semiconductor transition region having a spatially varying bandgap between the first and second interfaces to eliminate an abrupt change in bandgap between the multiplication region and the absorption region.
In yet another aspect, apparatuses and systems can include a semiconductor absorption region structured to receive light at a first wavelength to generate one or more charged carriers by absorbing received light; a multiplication region structured to receive the one or more charged carriers generated from the semiconductor absorption region and to generate an avalanche of secondary charged carriers in response to the one or more charged carriers and emit secondary photons from the secondary charged carriers; a region coupled between the absorption region and the multiplication region, the region comprising a mechanism that quenches the avalanche of the multiplication region after occurrence of the avalanche and resets the multiplication region for a next avalanche; and a semiconductor transition region formed between the buffer region and the absorption region to have a first bandgap at a first interface with the buffer region that is equal to or similar to a bandgap of the buffer region and a second bandgap at a second interface with the absorption region that is equal to or similar to a bandgap of the absorption region, the semiconductor transition region having a spatially varying bandgap between the first and second interfaces to eliminate an abrupt change in bandgap between the buffer region and the absorption region.
In yet another aspect, apparatuses and systems can include a semiconductor absorption region structured to absorb photons at a first wavelength to generate one or more charged carriers; a multiplication region structured to receive the one or more charged carriers, the multiplication region structured to generate an avalanche of electrons in response to the one or more charged carriers and emit secondary photons at a second wavelength shorter than the first wavelength; a buffer region structured to impede electrons or holes from the avalanche from passing through the buffer region to cause a reduction in an electric field across the multiplication region to quench the avalanche; and a bandgap grading region adjacent to the absorption region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the absorption region and a second region.
These and other implementations can include one or more of the following features. In some implementations, the buffer region can be structured to allow electrons to pass through the buffer region to cause an increase in the electric field across the multiplication region to facilitate a recovery from the avalanche. In some implementations, the buffer region can be structured to allow holes to pass through the buffer region to cause an increase in the electric field across the multiplication region to facilitate a recovery from the avalanche. In some implementations, the bandgap grading region can be positioned between the absorption region and the multiplication region, wherein the second region of the bandgap grading region interfaces with the multiplication region. In some implementations, the bandgap grading region can be positioned between the absorption region and the buffer region, wherein the second region of the bandgap grading region interfaces with the buffer region. The buffer region and the multiplication region can be structured to create an energy barrier by a valence band offset. Implementations can include a mechanism to bias the second optical sensing structure at a DC voltage. The multiplication region can be optically coupled with an optical sensing structure. The buffer region can include an InAlAs layer. The buffer region can include an InAlAs, InGaAsP, and InAlAs stack. The bandgap grading region can include one or more InGaAsP graded index layers.
The details of one or more implementations are set forth in the accompanying drawings, and the description and the claims in this document.
Like reference symbols in the various drawings indicate like elements.
This document describes, in one aspect, examples of single photon detection apparatuses and systems such as integrated infrared (IR) detectors and detector arrays based on a hybrid structure comprised of two avalanche detectors for detecting light at two different wavelengths that are coupled to each other in an integrated package. For example, an IR avalanche diode can be coupled to a silicon single-photon avalanche detector to form this hybrid structure. Such a device can be designed to produce electroluminescence resulting from hot-carrier recombination as a upconversion mechanism to convert received light into light of a shorter wavelength. The two detectors can be two SPADs.
In some implementations of the hybrid structure, an interconnection between an IR avalanche diode and the silicon avalanche diode is exclusively optically, and thus no electrical interconnection is required between the IR avalanche diode and the silicon avalanche diode. This optical interconnection makes the manufacturing of the devices simple and inexpensive because the two detectors formed on two different substrates can be coupled to each other by fusing their dielectric passivation layers without requiring lattice matching between the two devices. Advantageously, the hybrid structure can be made for various IR detecting materials at various IR wavelengths for the IR detector without limitations imposed by the lattice matching between the two devices. Because of lack of electrical interconnection, electrically conductive bumps for electrical interconnection are not needed and thus detector pixels in a detector array can be packed more closely together, making it possible to achieve small pixel pitch for high-resolution IR imaging.
The above hybrid structure of an IR detector and a Si detector is an example of a detection device or an array element that includes a primary detector such as an IR detector and a secondary detector that detects light emitted by the primary detector. The IR detector (the primary detector) may be formed of III-V materials (e.g., InGaAs/InP) and other suitable IR absorbing semiconductor materials. The IR detector may be structured to include a separate absorption layer and a multiplication layer in a separate absorption and multiplication (SAM) structure. The secondary detector, e.g., a silicon detector such as an shallow-trench-isolation (STI)-bounded Si detector or a suitable detector, can be based on a CMOS process, a BiCMOS process or other suitable silicon process.
In some implementations of the SAM structure, an emission can be increased by introducing an energy discontinuity at the edge of the multiplication region, e.g., away from the absorption layer. This discontinuity causes a local accumulation of the hot carriers, thereby increasing the probability of recombination and thus the electroluminescence yield that can be measured by the number of photons emitted per hot carrier. As an example, a semiconductor device based on the technical features in this document can include a substrate; a first optical sensing structure formed on the substrate to absorb light at a first optical wavelength; and a second optical sensing structure formed on a second substrate and fused over the first optical sensing structure. The second optical sensing structure can be structured to absorb light at a second optical wavelength longer than the first optical wavelength and to emit light at the first optical wavelength which is absorbed by the first optical sensing structure. The first and second optical wavelengths are different optical wavelength bands, e.g., a silicon detector can be used to detect light from 400 nm to 1100 nm. In one implementation of this device, the substrate is a silicon substrate, the first optical sensing structure is a silicon-based optical detector, and the second optical sensing structure is an IR optical detector which emits light at the first optical wavelength by converting energy in absorbed light at the second optical wavelength via luminescence resulting from hot carrier recombination. The second optical sensing structure can include an absorption structure which absorbs light at the second optical wavelength and a multiplication material layer between the absorption structure and the first optical sensing structure to emit light at the first optical wavelength. The absorption structure can have a bandgap less than a bandgap of the multiplication structure.
In implementations of detector devices based on the above hybrid structure, a detector device can include an IR avalanche detector and an Si detector that are fused through a silicon dioxide layer. The IR avalanche photodiode can be operated either in sub-Geiger or in Geiger mode as a single-photon detector. For example, the IR single photon avalanche photodiode (SPAD) can be constructed with a separate absorption and multiplication (SAM) zones. As a specific example, the absorption zone can be in InGaAs with a narrow bandgap, which can efficiently absorb photons, e.g., at 1550 nm which is used in communications and other applications, and the multiplication zone can be in InP, a direct bandgap material with a bandgap of 1.4 eV which is higher than that of silicon. A photogenerated electron-hole pair can induce an avalanche by a chain reaction of impact ionizations in the InP region. The accelerated (hot) electrons can impact stationary holes and the recombination of the hot electrons and the holes releases the excess energy with a known distribution that has a component above the energy bandgap of the material (InP). Therefore, this component manifests itself as a secondary radiation of electroluminescent photons and can be detected in silicon. The electroluminescent photons can be emitted isotropically. Some of such electroluminescent photons can reach the depletion region of the silicon SPAD, which is biased above its breakdown voltage. As a result, an avalanche breakdown occurs, and is detected by sensing circuitry, either on or off the same silicon die.
The IR and visible-wavelength SPADs in hybrid structures can be structured and operated in various configurations. For example, the IR SPAD can be configured to be back-illuminated so the emission occurs at the side facing the Si SPAD, or illumination can be through the back side of the silicon SPAD (which is transparent to IR wavelengths) and the primary photons in IR impinge directly on the front surface of the IR detector. The IR SPAD can have a multiplication region that includes a direct-bandgap material such as InP or GaAs, which has a higher photon-to-hot-electrons electroluminescence yield than indirect-bandgap materials. The IR SPAD can be configured to have its quenching mechanism contained within the device itself to maximize the performance gain from this scheme. The IR SPAD can have as shallow a junction as possible underneath the surface to minimize self-absorption of the secondary photons, and should have as thin a passivation layer as possible, to minimize dispersion of these photons away from the silicon SPAD. The silicon SPAD can be designed to have as wide an absorption region as possible, in order to increase the absorption probability of the electroluminescent photons. The two SPADs can be fused at the silicon dioxide layer or a suitable common passivation layer with minimal voids to reduce optical loss and enhance the mechanical integrity. The fusing can be performed either on wafer level or per die. The former can offer significant cost saving.
In addition, structures may be included in the two SPADs to enhance or maximize coupling between the two SPADs or to reduce inter-pixel cross-talk in an array of sensors. For example, such structures can include micro-lenses between the two SPAD detectors for collecting and focusing emitted light from the IR SPAD onto the silicon junction of the silicon SPAD; trenches between IR pixels to reduce optical cross-talk; metal rings around the silicon or IR pixels to prevent photons emitted from one pixel to reach adjacent pixels; and dielectric structures around the IR devices to reflect photons that are emitted away from the silicon device back towards the silicon device for efficient collection and detection of photons at the silicon device.
In some implementations, IR detectors tend to exhibit high dark counts which compromise the IR detection performance. Techniques are provided to significantly reduce or virtually eliminate the effect on the IR detection performance caused by the dark current counts. In cases where active illumination is used by, e.g., using a laser beam to illuminate a target under imaging, an electrical signal correlated to the laser pulse either biases the IR detector such that the IR detector is only active in a limited time window, or does the same to the silicon SPAD, or activates a read-out circuit such that signals are only read in a time window when signals are expected. Another scheme for eliminating the effect of dark counts can be used when the resolution of the image is less than the resolution of the IR array. In such a scenario, a “real” illumination will cause an avalanche in at least two adjacent IR pixels, which will be detected by a corresponding number of silicon pixels. A logic gate can locally process the SPAD outputs and only pass those pulses which are valid (dark counts are non-correlated so the probability of two adjacent cells firing simultaneously due to dark counts can be shown to be negligible for realistic sampling time-gates).
Such an integrated detector structure can be used to provide scalability for forming large arrays such as processing parallel communication channels simultaneously and array imaging (e.g., 3D imaging), as well as for high-sensitivity infrared imaging. The pitch of detection pixels in this case is not limited by present wafer-level interconnection schemes such as indium bumps or cross-vias because optical interconnection is used. Optical and/or physical isolation structures may be used, such as mesa structures, for limiting the avalanche propagation or metal barriers to provide optical isolation between adjacent pixels. In some implementations, an electrical readout can be eliminated for a primary detector to reduce the junction capacitance of an IR detector, because a secondary detector can process optical information coming from the primary detector. The described technologies can produce significant reductions in power consumption and other performance improvements such as improved speed and reduced afterpulsing.
Several techniques can be used to evaluate single-photon detectors. The single-photon detection probability, η, is the product of the probabilities of a photon being absorbed in the material and of it initiating a detectable avalanche. The detector's spectral response describes the wavelength-dependence of this detection probability. These metrics depend on the percentage of pixel area which collect photons (fill ratio); on the absorbing layer's composition, depth and thickness; and on the electric field distribution in the multiplication region.
During the recharge process following an avalanche, the SPAD can be temporarily biased below the breakdown voltage, and cannot generate an avalanche pulse in response to a photon. This time is called the device dead time and depends on the recharge mechanism, on the overbias above breakdown and, most significantly, on the junction's capacitance.
Dark counts result from avalanches which are not induced by absorbed photons. They can originate from thermally-generated carriers; from band-to-band tunneling; via trap-assisted tunneling; and by afterpulsing—the release of carriers trapped in prior avalanches. The latter mechanism is an important factor in determining the device dead time.
The time-delay spread between the photon absorption event and the clocking of the resulting electrical signal depends on the diameter of the SPAD as well as on the timing circuitry. The time-delay spread can determine the timing resolution of the single-photon detector.
Solid-state IR SPADs detectors can have separate absorption and multiplication regions, whereby, for example, photons are absorbed in a structure such as a thick (e.g., several microns) lightly-doped InGaAs layer. The photo-generated carriers are swept towards an InP high-field multiplication region where impact ionizations provide gain. When the extraction rate of carriers from this multiplication region falls below the creation rate, an avalanche breakdown is said to occur and the gain becomes “infinite”. Some detectors can operate in a Geiger Mode (GM) for single-photon detection. The avalanche must be quenched to avoid damage to the junction.
In some implementations, quenching can be achieved passively, by using a voltage-limiting resistor in series with the device, or actively, using a circuit which senses the onset of the avalanche, and subsequently quenches it. Once the avalanche has been quenched, the diode capacitance can be re-charged, either passively, through the quenching resistor, or actively, using a recharging circuit. In some implementations, a SPAD can be self-quenched.
IR Geiger-mode single photon avalanche diodes (GM-SPADs) can be amenable to integration in large arrays and to mass production because they can be manufactured using standard lithographically-defined processing techniques. However, the support circuitry, including the quenching, recharging and processing circuitry must be implemented externally, usually in silicon.
Further, GM-SPADs may experience excessive amounts of noise. Hole-electron pairs, thermally generated at the edge of the high field region through Shockley-Read-Hall generation and separated by the strong electric field, can cause a “false” avalanche. Trap-assisted tunneling depends on the defect density in addition to the doping and may be exacerbated at high electric fields by barrier lowering via the Poole-Frenkel effect. Direct band-to-band tunneling requires strong electric fields above 7×105 V/cm occurs in devices with a breakdown voltage lower than 4EG/q, where EG is the bandgap energy and q is the electron charge. Finally, afterpulsing which results from the release of deep traps trapped during previous SPAD cycles, increases with high defect densities and is linearly dependent on the total charge flowing during an avalanche. The rate of emission of deep traps follows an exponentially decaying distribution which depends on the activation energy of each deep trap mechanism.
Some of these noise sources can be reduced by cooling. Deep trap lifetimes can increase exponentially as temperature is decreased. Thermal generation and tunneling increase as the bandgap decreases. Thus, in some implementations, cooling an IR SPAD may be required. For example, some IR SPAD may be cooled to about 200° K. At these temperatures, afterpulsing becomes the dominant noise source with rates on the order of tens of kHz and it becomes the main bottleneck for device bandwidth. Time-gating can reduce the effect of afterpulsing but due to its exponential time distribution, the separation between gates (device dead time) must be made long enough compared with the afterpulse lifetime, in order to sufficiently reduce the probability of experiencing an afterpulse during an exposure time gate. Furthermore, time gating is possible when the arrival time of the photon is known to within the duration of the gate. At low temperatures, SPADs can be operated in free-running mode with minimal afterpulsing effects where a sufficiently long hold-off time is used following an avalanche, thereby severely limiting the detection rate.
For a given technology, afterpulsing can be reduced by limiting the charge flowing during an avalanche. This may be done by active quenching but is achieved more efficiently by reducing the junction capacitance. The capacitance in IR SPADs is dominated by the capacitances of the readout, recharge and quenching circuitry, because these operations are implemented off-chip, either on a board or on a silicon die. Connections can be made using technologies such as wire bonding or by using indium bumps. This can limit the pixel pitch and may result in capacitances on the order of pF, thus deteriorating the noise performance of the device due to afterpulsing.
This document includes a description of integrated detectors that include a new interconnection and readout scheme which does not require electrical interconnections between an IR SPAD and a CMOS readout circuitry, offers superior upconversion efficiencies with low power, and is scalable to large arrays. By bypassing the requirement of electrical bonding between the SPAD pixels and the readout circuitry, the capacitance seen by the junction is significantly reduced. This results in a reduction in afterpulsing, which is the dominant noise source at low temperatures, while simultaneously decreasing the detector device's dead time.
An integrated detector device can be based on wavelength upconversion using a byproduct of the avalanche process, e.g., hot-carrier luminescence from a multiplication layer of the device. This luminescence can have a significant component at higher energies than the bandgap of the absorbing material, and can therefore be extended to several IR detection materials. Readout of the luminescent photons can be achieved by a coupled detector such as a silicon single-photon avalanche diode.
A manufacturing process for an integrated detector device can include utilizing a mature wafer-level glass-to-glass fusing technology in order to connect different components of an integrated detector device (see, e.g., J. Wei, S. M. L. Nai, C. K. S. Wong, Z. Sun, and L. C. Lee, “Low temperature glass-to-glass wafer bonding,” IEEE Transactions on Advanced Packaging, vol. 26, pp. 289-294, 2003). For example, a wafer-level glass-to-glass fusing technology can form a connection between a III-V SPAD which detects the primary IR photon, and a silicon CMOS SPAD, which detects the up-converted photons, and which processes the information on the same die.
Hot-carrier luminescence can result from a recombination of hot electrons with holes. In direct bandgap materials, a photon is emitted whose energy equals the difference between the hot electron's initial energy and the bandgap. In an indirect recombination process, both energy and momentum must be exchanged, and the probability of such events is considerably lower, depending on the availability of suitable phonons. This is manifested by the different electron temperatures in these processes. In GaAs, the direct recombination process is characterized by an electron temperature of 800K while the indirect process in the same material exhibits a temperature of 3000K. The higher temperature stems from a longer mean lifetime, consistent with a less probable recombination event.
ω>Egd. A low-energy, infrared, component due to transitions between the light and heavy hole bands is also observed.
For a direct bandgap material, such as GaAs or InP, the photon emission rate is given by:
Rd(ω)∂
ω(
ω−Egd)1/2f(E)[1−f(E−
ω)] (1)
where ω is the emitted photons' energy, Egd is the direct bandgap; E is the electron energy above the bottom of the conduction band; f(E) and [1−f(E−
ω)] are the hot electron and the hole distributions, respectively, both of which strongly depend on the hot-carrier temperature.
ω>Egd thereby achieving energy upconversion.
The efficiency of this upconversion depends on the electroluminescence yield, e.g., the photon emission rate per unit avalanche charge. This figure can be difficult to measure due to self-absorption by the emitting device, the detector's spectral response, the effect of defects, and collection uncertainties due to reflections. Kurtsiefer et al reported a figure of 39 photons per steradian in an avalanche with 4×108 electrons, resulting in a lower limit of 2.5×10−6 photons per electron, where the detector's spectral response has been partly accounted for, and self-absorption was not (C. Kurtsiefer, P. Zarda, S. Mayer, and H. Weinfurter, “The breakdown flash of silicon avalanche photodiodes-back door for eavesdropper attacks?,” Journal of Modern Optics, vol. 48, pp. 2039-2047, 2001). A measurement accounting for both the optical system and self-absorption was presented by Lacaita, with an emission efficiency of 2.9×10−5 photons with energy higher than 1.14 eV per carrier crossing the junction (A. L. Lacaita, F. Zappa, S. Bigliardi, and M. Manfredi, “On the Bremsstrahlung origin of hot-carrier-induced photons in silicon devices,” IEEE Trans. Electron Dev., vol. 40, pp. 577-582, 1993). Electroluminescence yield for InP has not been reported to date. It is expected to be significantly higher than that of silicon and is conservatively estimated to be 2.9×10−4 photons per hot-carrier for the purposes of our calculation.
In some implementations, an InGaAs SPAD is direct-bonded to a Si SPAD for optical readout. IR photons are incident on the back surface of the InGaAs SPAD. As carriers recombine during the avalanche, they release NIR and visible photons which are detected by the silicon device. The devices can be fused through their silicon-dioxide passivation layer, eliminating the need for lattice matching between the two semiconducting materials.
Determining the overall detection probability of the up-conversion scheme can include multiplying the primary NIR detection probability in the InGaAs/InP SPAD, by the emission probabilities (as a function of wavelength) of the electroluminescent photons emitted towards the silicon SPAD. One can account for self-absorption in the InP device. One can determine the probability of absorption of these photons in the silicon SPAD's depletion region, and can calculate the avalanche initiation probability of the photogenerated charge carriers. In some implementations, an up conversion probability can be expressed as:
where Ω is the solid angle subtended by the silicon junction when seen from the InP junction; Nsp is the number of electroluminescent photons at energies, ω, emitted in a primary avalanche; Psa(
ω,xj1) is the probability of self-absorption of the secondary photons, which is a function of their energy and generation depth xj1; Pabs is the absorption probability in the silicon SPAD's depletion region, which extends from xj2 to xj2+xd; and Pav, (xabs) is the probability for an electron-hole pair photo-generated at xabs to induce a detectable avalanche.
Operations of an integrated IR detector device can cause a secondary photon emission towards a silicon junction in response to one or more photons hitting the primary detector of the device. An avalanche event in a SPAD can be viewed as a discharge of the junction capacitance, Cj, from an initial voltage, in excess of the diode's breakdown voltage to approximately the breakdown voltage. The total number of electrons flowing during an avalanche is:
where q is the electron charge (in Coulomb) Cj is the junction capacitance, Cp is any additional capacitance seen by the junction, including interconnection and sensing capacitances and Vob is the overbias above breakdown.
The number of secondary photons emitted from the primary junction can be expressed as:
Nsp(ω)=ηes(
ω)Ne (4)
where ηe is the luminescence yield per electron (in relevant energies for Si absorption) and s(ω) is the normalized spectral distribution of the secondary photons. In some implementations, additional photons can be expected to be emitted from the charge region between the absorption and multiplication regions, where the high electric field is lower than the breakdown field, and thus recombination events are highly likely. It can be assumed that secondary photons are emitted from the maximum field region, at the junction plane.
is transmitted towards the silicon junction. The solid angle, Ω, for the case of a planar InP junction emitting towards a parallel planar silicon junction can be approximated by assuming all photons are emitted from one of the vertices of the InP rectangular junction, e.g.,
where d is the side dimension of the Si junction and S is the vertical distance between the junctions. Equations (3), (4), and (5) can be combined to give:
Self-absorption in InP reduces the number of photons which reach the surface. It can be assumed that luminescence occurs at the junction plane, a distance xj1 from the surface. The emitted photon population can be expressed as:
where αInP (ω) is an absorption coefficient in InP.
Based on a calculated spectral distribution of the emitted photons, one can estimate the probability for these photons to be absorbed by the Si SPAD, and can estimate the probability of generating an avalanche. It can be assumed that reflections at the interfaces do not substantially affect the number and spectral distribution of secondary photons.
In an STI-bounded shallow junction, the high field region can be highly localized in the depletion region of the junction, so it can be assumed that absorption occurs within the depletion region. The probability for N photons of energy ω to generate an electron-hole pair within this layer is:
PabsN(ω)=1−{1−Pabs(exp[−αSi(
ω)·wd]−exp[−αSi(
ω)·(xj2+wd)])}N (8)
with αSi being the absorption coefficient in silicon, wd the depletion width and xj2 the junction depth in the silicon device. The depletion width of the junction can be determined from the analytical expression for a one-sided linearly-graded junction:
where VB is the sum of applied and built-in voltages, ∈s is the dielectric constant of silicon, q the electron charge and a the grading coefficient of the linearly-graded junction. The total upconverted photons' absorption probability in silicon can be calculated using equations (7), (8), and (9) over relevant wavelengths.
The probability that an absorbed photon induces an avalanche can be calculated. For a one-sided, linearly-graded pn junction, Poisson's equation translates to a field distribution:
The avalanche probability can be estimated as a function of the position of generation of the electron-hole pair by solving the following coupled differential equations:
where Pbe and Pbh are the avalanche initiation probability by an electron and a hole, respectively, α and β are the ionization rates of electrons and holes, respectively, and Pbp is the joint avalanche initiation probability:
Pbp=1−(1−Pbe)(1−Pbh)=Pbe+Pbh−PbePbh (12)
These equations can be solved numerically to provide an avalanching probability.
In some implementations, a detector device can include a self-quenched InGaAs/InP SPAD with an active area of 15 μm per side and can have a junction capacitance of 150 fF, dominated by the capacitance of the depleted region. Due to the optical readout, off-chip routing and the sensing circuit's capacitance, which can be on the order of a picofarad in SPADs with electrical readout, are eliminated. In some implementations, the passivation thickness of a 6-metal layer silicon device is on the order of 7 μm, so InP-SPAD/Si-SPAD capacitance can be negligible. In some implementations, a self-quenched InGaAs/InP SPAD can operate at an overbias of 5V with a junction located 200 nm below the surface.
A detector device, such as a device with primary and secondary detectors, can include a mechanism to prevent a positive feedback loop between the detectors. A silicon SPAD, positioned as a secondary detector, can emit electroluminescent photons which can be absorded by an IR SPAD, positioned as a primary detector. Controlling the dead time of the IR SPAD to overlap the avalanche time of a silicon SPAD can prevent a positive feedback loop between the coupled SPADs.
In some implementations, the power dissipated during an up conversion process is the sum of the powers dissipated during the InP and silicon avalanche. These can be estimated as the product of the junction capacitances by their overbias, resulting in approximately 1 pW per detected photon.
Various examples of photon based systems that include one or more detectors can include systems for quantum key distribution (QKD), quantum communications, pulse position modulation communication links, fluorescence lifetime imaging microscopy (FLIM), and time-correlated single-photon counting (TCSPC), eye-safe laser detection and ranging (LIDAR), optical time-domain reflectometry (OTDR) and semiconductor failure analysis systems. In some implementations, single-photon detectors can operate at high frequencies such as tens to hundreds of MHz, can consume minimal power (<1 nW/bit), and can operate reliably at non-cryogenic temperatures over many cycles and can be manufactured at a low cost. When operated in arrays, such detector devices can have a small pitch and low pixel-to-pixel cross-talk. Further, such systems can feature low detector jitter for more efficient operations. For example, lower detector jitter can result in lower bit-error-rates in quantum key distribution systems, faster bit-rates in pulse position modulation optical links, and better temporal resolution in biological applications.
A silicon SPAD, such as a CMOS SPAD, can include a diffused guard-ring structure which can lead to increased jitter due to the lateral drift and diffusion of electron-hole pairs created by photon absorption in the low-field and neutral regions. The resulting diffusion tail can limit SPAD timing performance, e.g., in QKD systems and in classical pulse position modulation optical links, where diffusion tails limit bit error rates and bit rates, respectively, and in high-resolution FLIM, where better timing resolution translates into better image contrast. Therefore, in some implementations, a silicon SPAD can feature a SPAD architecture optimized not only for a minimal FWHM but also for a minimal full-width at hundredth-maximum (FW( 1/100)M).
A silicon SPAD, such as a CMOS SPAD, that has a low detector jitter suitable for implementations of a detection device that integrates a silicon SPAD and an IR SPAD can include a shallow-trench-isolation (STI) guard ring structure to reduce jitter and to reduce or eliminate a diffusion tail. An area-efficient shallow-trench-isolation guard ring structure can prevent lateral drift and diffusion of charge carriers to enhance the detector's timing resolution. Certain technical information for silicon STI-SPADS is included in PCT Application No. PCT/US2007/074057 entitled “Shallow-Trench-Isolation (STI)-Bounded Single-Photon Avalanche Photodetectors” and published as PCT publication No. WO 2008011617, which is incorporated by reference in its entirety as part of the disclosure of this document. The trench isolation guard ring is capable of withstanding high electric fields and encloses a boundary of a p-n junction region to spatially confine diffusion of charges at the p-n junction to planarize the interface of the p-n junction of the diode. As a result, the guard ring can be used to prevent premature breakdown and to enhance uniformity of the electric filed distribution along the pn junction and thus the detection probability across the detection area or active area of the p-n junction. A guard ring can be formed to enclose the boundary of either or both of the shallow and deep pn junctions. Implementations of such guard rings can be used to achieve high fill factors and small pixels in compact and high performance sensor arrays.
A STI-based CMOS SPAD fabrication process can include a deep-submicrometer CMOS process such as one based on 0.18-μm CMOS technology. A STI's isolation trench can be constructed early in the fabrication process. In some cases, an isolation trench can prevent punch-through and latch-up in CMOS circuits. The edges of the subsequent drain implant are confined by an oxide trench that prevents lateral diffusion and formation of curved edges. In some implementations, such a manufacturing process can include using an IBM 0.18-μm CMOS technology through the MOSIS service, packaged in a high-speed quad flat no-lead (QFN) glass-top package with ultrashort wiring to minimize capacitance and inductance.
In some implementations, two different output buffers can be used, e.g., a source follower and an inverter chain. A source-follower buffer can use a circuit that outputs a voltage proportional to the current flow to allow a direct observation of an avalanche pulse. An inverter chain output buffer can generate uniform-amplitude pulses for improved compatibility with pulse counting electronics.
In addition to the low jitter of the STI-bound SPAD device, such as device can exhibit shorter dead times and higher fill factor than diffused guard-ring SPADs. This can come at the expense of relatively high dark count rates (e.g., 104-106 counts/s) and afterpulsing, possibly due to interface states at the SiO2—silicon boundary. Device implementations can include time gating and active recharging to reduce the effects of this noise to acceptable levels. In some implementations, a STI-bound SPAD can be scaled to megapixel arrays to enable improved performance in optical communication channels and better temporal resolution in biological applications.
A detector device can include an array of detector pixels. In some implementations, an array of III-V detector pixels can be constructed such that an avalanche region is spatially confined and an array of Si SPADs is aligned (e.g., using infrared imaging since it is transparent to Si) to the III-V detector pixels such that the active areas of the Si SPAD is aligned to the avalanche regions of the III-V semiconductor region. In some implementations, a detector device can include a metal layer on the III-V semiconductor region to minimize inter-pixel cross-talk. In some implementations, a detector device can include a metal layer on the Si semiconductor region semiconductor region to minimize inter-pixel cross-talk. A detection system can include an array of detector pixels configured for low-light level infrared imaging.
A detector device can be configured for front side illumination or back side illumination. In a front side illumination example, IR photons impinge on the back surface (bulk side) of the silicon wafer, pass through it and the SiO2 dielectric (since Si is transparent to IR photons) and is absorbed by the III-V junction. In a back side illumination scheme, the III-V junction can be formed on top of an IR-transparent substrate. IR photons impinge on the back side of the III-V detector.
Some photon detector devices can include a mechanism for self-quenching and self-recovery within an InGaAs SPAD. At a critical field, the high probability of impact ionization through a semiconductor can generate a large avalanche current. The multiplication gain resulting from a single carrier undergoing this process can be in the millions, so a single photogenerated carrier can be detected. Left alone, this large avalanche current can continue and, due to presence of the large avalanche current caused by the first photon, the photodetector cannot detect a second photon. Some InGaAs SPADs cannot self-quench or cannot self-recover because they use an external circuit for quenching and recovery. However, such external circuitry can increase manufacturing costs and have performance issues. This document includes details of SPADs, such as an InGaAs SPAD, with a self-quenching and self-recovery mechanism without using the external quenching and recovery circuitry. For example, a SPAD can include a semiconductor layer in a detector stack that functions as a negative feedback mechanism for self-quenching and self-recovery.
A self-quenching, self-recovering SPAD can be configured to produce secondary photon emission. In some implementations, a self quenching avalanche photodiode can be made via bandgap engineering by using III-V materials latticed matched to InP. A buffer region such as a Transit Carrier Buffer (TCB) can be positioned next to a multiplication region to generate an energy barrier. Then the resulting electrons from the avalanche will be momentarily stopped by the barrier, e.g., TCB. This barrier can reduce the field across the multiplication region and thus stops the avalanche process. When the avalanche process triggered by an incident photon is stopped, the device is self quenched. As the electrons escape from the barrier, the field across the multiplication region recovers. At this time, the self-quenched detector is capable of detecting a second photon via another avalanche process.
The high energies of carriers undergoing avalanche breakdown can result in photoemission with energies larger than the bandgap of the material. By utilizing this property, a separate absorption and multiplication APD can detect low energy photons with a small bandgap absorption region, and generate photons in the multiplication region with higher energies. For example, an InGaAs—InP system can detect 1550 nm infrared, and output light such as visible light. In some implementations, the multiplication region is placed near the surface of the device to minimize self-absorption.
In some implementations, a self-quenching SPAD can include a self quenching layer, such as a TCB, designed by using the band offsets of various epitaxial layers lattice matched to an InP substrate. A region with a high band offset, called a transit carrier buffer (TCB), is placed next to a multiplication region. Carriers generated from the avalanche collect at the interface due to the band offset of the heterojunction between the TCB and the multiplication region, decreasing the electric field through the multiplication region below breakdown in a short time (e.g., <0.1 ns), self-quenching the device. After the avalanche pulse, on a longer time scale (e.g., 10 to 100 ns), the carriers escape from the barrier via thermal excitation and tunneling, self-recovering the device to a ready state for detecting another photon.
A TCB can include a material with a valence band offset next to the p-layer of the multiplication p-i-n to stop holes. In some implementations, a TCB can include material with a conduction band offset next to an n-layer of p-i-n semiconductor layers to stop electrons.
In some implementations, by coupling a negative feedback mechanism with the gain generated from the avalanche process, the current and voltage response of the device can then be simulated in Geiger mode operation, although the DC gain in this mode is infinite.
Referring to
Upon reaching x=0, additional carriers are generated through impact ionization. A deterministic Selberherr's model is used, with impact ionization dependent only on the local field near the carrier. The ionization rates for electrons and holes are then given by equations (13) and (14), respectively.
α(E)=α0exp(−[cn/E]m
β(E)=β0exp(−[cp/E]m
The continuity equations in the multiplication region are
where the generation rates are Gp=Gn=βJp+αJn. The current in the multiplication region is then given by
where it is assumed the carriers travel at their saturation velocity due to the high field. The total current is the summation of the particle current and the displacement current:
where Vm is the voltage across the multiplication region. Since the current J is independent of x, the following equation is derived:
The voltage applied across the device is assumed to drop only due to the multiplication and TCB regions. Then the voltage across the TCB region is VTCB=V−Vm. As the avalanche process builds up, holes are trapped at the TCB interface due to the band discontinuity, forming a sheet charge Qi. The TCB region can be modeled as a leaky capacitor, e.g., a capacitor and resistor in parallel. This would have a time constant τ, which represents the escape time of holes from the interface. The current through the TCB region is then
Using the boundary condition of zero electrons at the end of the multiplication region (e.g., at x=W), with the definitions T=τ(1+L/W), A=β−α, and ν=2(νp+νn), the above equations can be rearranged into
A numerical solution to equation (22) can be calculated with boundary conditions Vm=Vapplied and dVm/dt=0 for t<0. Given an effective cross sectional area S (typically about 100 μm2), a single photon response can be modeled by
Jp(0,t)=δ(t)e/S (23)
Unless noted otherwise, calculations are made at a 5% relative overbias, e.g., 105% of the breakdown voltage. The breakdown voltage is determined by setting the multiplication factor to infinity in equation (24), where k=β/α and the ionization coefficients are assumed to be position independent in the p-i-n junction.
The dependence of several device characteristics on the applied bias can be found from the current response of a single photon. The gain is calculated by integrating the current under the avalanche pulse, and is shown to have a linear relationship to the applied bias (e.g., see
The current response to an input photon is strongly dependent on the voltage across the multiplication region. If a secondary photon is introduced before full recovery, the resulting current gain may be less than ideal, and thus the device can have a maximum detection rate depending on the output sensitivity needed.
This same setup can also be used to model the dark current levels of the device. If it is assumed dark carriers are generated at a given rate, the dependence of the average current level on the applied bias can be plotted (e.g., see
To optimize device characteristics such as gain and recovery time, the variation of the device structure and material parameters is examined. The hole escape time τ, which physically depends on barrier height and thickness, along with the TCB layer thickness L is varied. Longer hole escape times lead to a longer self-recovery time of the device. In some implementations, shorter escape times mean there will be fewer holes trapped at the interface, leading to a longer quenching time. As the device takes longer to quench, the device can remain in a Geiger mode for a longer time, and thus produces a larger gain.
The modeling of the SPAD with negative feedback shows gain can be increased by increasing the applied overbias, at the cost of increasing dark current. Increasing the applied bias will also benefit key characteristics such as the quenching time. To maximize gain, either the self-quenching barrier height (related to hole escape time) or barrier thickness can be modified. As either is decreased, the infinite gain of a conventional SPAD in Geiger mode is recovered; but doing so diminishes the fast self-quenching property. As gain is much more sensitive to the barrier thickness, a thin and high barrier is preferred to a low and thick barrier.
A detector can include an InGaAs/InAlAs based SPAD to detect single photons at λ=1550 nm, with a quenching layer InGaAsP (Eg=1 μm). In such a detector, the resulting valence band offset between an InAlAs multiplication region and a TCB region is around 80 meV.
The avalanche pulse width is found to be around 30 ns (limited by the electronics rather than the intrinsic response), with a device recovery time of around 300 ns. Due to the self-quenching mechanism, the avalanche pulses can be uniform.
In the SPAD design in
In some implementations, a detector device can include a semiconductor absorption region structured to absorb photons at a first wavelength to generate one or more charged carriers. The detector device can include a multiplication region structured to receive the one or more charged carriers, the multiplication region structured to generate an avalanche of electrons in response to the one or more charged carriers and emit secondary photons at a second wavelength shorter than the first wavelength. The detector device can include a buffer region structured to impede electrons or holes from the avalanche from passing through the buffer region to cause a reduction in an electric field across the multiplication region to quench the avalanche. The detector device can include a bandgap grading region adjacent to the absorption region, at least a portion of the bandgap grading region having a spatially varying bandgap profile that monotonically changes between a first region that interfaces with the absorption region and a second region.
Different TCB structures and arrangements can be implemented. Table 1 and Table 2 show different example of detector device structures.
Table 1 shows an example of an epitaxial layer structure of a self-quenching and self-recovering device. An InAlAs layer is placed in between the InP multiplication and InGaAs absorption regions to act as a TCB for electrons. Graded transition layers are placed in between to support the transport of photogenerated holes from InGaAs to InP. The system can be capped with a thin p layer for an ohmic contact.
TABLE 1
Material
Thickness
Doping
Function
InGaAsP
0.1 μm
p = 1e18
Ohmic contact
InP
0.2 μm
p = 2e17
Multiplication region
InP
0.8 μm
Multiplication region
InP
0.25 μm
n = 1e17
Multiplication region
InGaAsP GRIN
0.1 μm
Transition layer
(1.1 um to InP)
InAlAs
0.8 μm
TCB
InAlGaAs GRIN
0.3 μm
Transition layer
InGaAs
1.5 μm
Absorption layer
InP buffer
0.5 μm
n = 1e17
InP n substrate
TABLE 2
Material
Thickness
Doping
Function
InGaAsP
0.1 μm
p = 1e18
Ohmic contact
InP
0.2 μm
p = 2e17
Multiplication region
InP
0.8 μm
Multiplication region
InGaAsP GRIN
0.1 μm
Transition layer
(1.1 um to InP)
InAlAs
0.3 μm
n = 7e16
TCB
InGaAsP (1.1 um)
0.2 μm
InAlAs
0.3 μm
TCB
InAlGaAs GRIN
0.15 μm
Transition layer
InGaAs
1.5 μm
Absorption layer
InP buffer
0.5 μm
n = 1e17
InP n substrate
Table 2 shows a different example of an epitaxial layer structure of a self-quenching and self-recovering device. In some implementations, a device can include two or more TCB layers. In some implementations, the field drop can occur across the first TCB barrier.
A detector device can include an InGaAs/InAlAs SPAD for near infrared light detection with build-in self-quenching and self-recovery capabilities. The device can detect light signal down to single photon level at near infrared wavelength. In some implementations, the detector device integrates four functions, photon absorption, avalanche multiplication, avalanche quenching, and device resetting. The build-in self-quenching self-recovery capabilities can enable the device to be operated with a simpler read-out circuit and in sub-Geiger mode, which is in contrast to conventional SPADs Geiger mode operation. In sub-Geiger mode operation, the device is biased at DC voltage. In some implementations, a detector device can include a mechanism, such an electronic circuit, to operate the InGaAs/InAlAs SPAD in a DC condition for sub-Geiger operations. Biasing the detector device at DC voltage can reduce detection system complexity and can minimize the SPAD's dead time. The detector device can achieve high multiplication gain with very low excess noise. The detector device can be configured to count photons for photon number resolving applications. The detector device can have a stable gain versus bias behavior.
An InGaAs/InAlAs SPAD can exhibit a gain saturation behavior, e.g., as increasing the bias, the gain can increase to a certain value, (˜105 to 106), and become saturated. The saturation gain value can be tuned by adjusting the energy barrier height and the quenching layer capacitance. A detection system can use a SPAD with this gain saturation property for increased reliability, and for a lower voltage supply requirement, which, in turn can reduce system cost.
An InGaAs/InAlAs SPAD based on the techniques described in this document can be configured for self-quenching. A self-quenching process can regulate an avalanche multiplication process, in a multiplication region of the SPAD, to achieve high gain (e.g., 106) for single photon detection and in the same time suppress noise to a low level.
In some implementations, a detector device for single photon frequency up-conversion can include an IR SPAD optically coupled with a CMOS SPAD. The photons produced by hot-carrier recombination process in the IR SPAD can be subsequently sensed by the CMOS SPAD to allow for on-die data processing. In some implementations, a CMOS SPAD can detect up-converted, visible photons and can include an on-chip readout circuitry to process the signal. Coupling between the IR and CMOS SPAD can be accomplished using mature wafer-level glass-to-glass fusing technology to increase manufacturing efficiency and to reduce costs when compared to other hetero-integration approaches. In some implementations, the IR SPAD and CMOS SPAD can be fused through a silicon dioxide passivation layer. To achieve up conversion in the detector device, the device can utilize a byproduct of the avalanche process, specifically the spectral component of the electroluminescent photons which is higher than the bandgap of the absorbing material of the detector. The detector device can exhibit low noise, can have a short dead time, can have superior up-conversion efficiencies with low power and can be scalable to large arrays.
Detector device 2200 can be structured to provide the absorption of IR and multiplication of the carriers to produce light at an optical wavelength shorter than the IR light. The absorption material can be a narrow bandgap IR absorption material and thus can absorb in the IR spectral range. The multiplication material can have a wide bandgap greater than the narrow bandgap to reduce noise. The electroluminescence peaks near the bandgap of the multiplication material's bandgap and thus has higher energy than the absorbed photon. It also has a tail with even higher energies, due to the energy distribution of hot carriers. This is how the upconversion is achieved. Even in materials with the same absorption and multiplication materials, upconversion is possible, due to the above-mentioned tail, but in this case it will be less efficient. Also, if the multiplication material is a direct bandgap material, electroluminescence through hot-carrier recombination can be more efficient.
While this document contains many specifics, these should not be construed as limitations on the scope of an invention or of what may be claimed, but rather as descriptions of features specific to particular embodiments of the invention. Certain features that are described in this document in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or a variation of a subcombination.
Only a few implementations are disclosed. However, variations and enhancements of the described implementations and other implementations can be made based on what is described and illustrated.
Finkelstein, Hod, Cheng, James, Lo, Yu-Hwa, Zhao, Kai, Esener, Sadik C., You, Sifang
Patent | Priority | Assignee | Title |
10263032, | Mar 04 2013 | Apple, Inc. | Photodiode with different electric potential regions for image sensors |
10285626, | Feb 14 2014 | Apple Inc. | Activity identification using an optical heart rate monitor |
10424683, | May 17 2018 | HI LLC | Photodetector comprising a single photon avalanche diode and a capacitor |
10438987, | Sep 23 2016 | Apple Inc | Stacked backside illuminated SPAD array |
10440301, | Sep 08 2017 | Apple Inc. | Image capture device, pixel, and method providing improved phase detection auto-focus performance |
10453881, | Nov 17 2015 | Taiwan Semiconductor Manufacturing Co., Ltd. | Infrared image sensor component |
10515993, | May 17 2018 | HI LLC | Stacked photodetector assemblies |
10529891, | Jul 31 2015 | International Business Machines Corporation | Resonant cavity strained III-V photodetector and LED on silicon substrate |
10559714, | Jul 31 2015 | International Business Machines Corporation | Resonant cavity strained III-V photodetector and LED on silicon substrate |
10609348, | May 30 2014 | Apple Inc. | Pixel binning in an image sensor |
10622538, | Jul 18 2017 | Apple Inc. | Techniques for providing a haptic output and sensing a haptic input using a piezoelectric body |
10656251, | Jan 25 2017 | Apple Inc | Signal acquisition in a SPAD detector |
10658419, | Sep 23 2016 | Apple Inc | Stacked backside illuminated SPAD array |
10672935, | May 17 2018 | HI LLC | Non-invasive wearable brain interface systems including a headgear and a plurality of self-contained photodetector units |
10672936, | May 17 2018 | HI LLC | Wearable systems with fast-gated photodetector architectures having a single photon avalanche diode and capacitor |
10797193, | Jan 23 2018 | Lumentum Operations LLC | Bias control structure for avalanche photodiodes |
10801886, | Jan 25 2017 | Apple Inc | SPAD detector having modulated sensitivity |
10847563, | May 17 2018 | HI LLC | Wearable systems with stacked photodetector assemblies |
10848693, | Jul 18 2018 | Apple Inc. | Image flare detection using asymmetric pixels |
10928492, | Jan 25 2017 | Apple Inc. | Management of histogram memory for a single-photon avalanche diode detector |
10943935, | Mar 06 2013 | Apple Inc. | Methods for transferring charge in an image sensor |
10962628, | Jan 26 2017 | Apple Inc | Spatial temporal weighting in a SPAD detector |
11004998, | May 17 2018 | HI LLC | Wearable brain interface systems including a headgear and a plurality of photodetector units |
11006876, | Dec 21 2018 | HI LLC | Biofeedback for awareness and modulation of mental state using a non-invasive brain interface system and method |
11019294, | Jul 18 2018 | Apple Inc | Seamless readout mode transitions in image sensors |
11069832, | Jul 31 2015 | International Business Machines Corporation | Resonant cavity strained III-V photodetector and LED on silicon substrate |
11069833, | Jul 31 2015 | International Business Machines Corporation | Resonant cavity strained III-V photodetector and LED on silicon substrate |
11096620, | Feb 21 2020 | HI LLC | Wearable module assemblies for an optical measurement system |
11187575, | Mar 20 2020 | HI LLC | High density optical measurement systems with minimal number of light sources |
11213206, | Jul 17 2018 | HI LLC | Non-invasive measurement systems with single-photon counting camera |
11213245, | Jun 20 2018 | HI LLC | Spatial and temporal-based diffusive correlation spectroscopy systems and methods |
11233966, | Nov 29 2018 | Apple Inc. | Breakdown voltage monitoring for avalanche diodes |
11245404, | Mar 20 2020 | HI LLC | Phase lock loop circuit based signal generation in an optical measurement system |
11271031, | Sep 23 2016 | Apple Inc. | Back-illuminated single-photon avalanche diode |
11398578, | Jun 06 2019 | HI LLC | Photodetector systems with low-power time-to-digital converter architectures to determine an arrival time of photon at a photodetector based on event detection time window |
11437538, | May 17 2018 | HI LLC | Wearable brain interface systems including a headgear and a plurality of photodetector units each housing a photodetector configured to be controlled by a master control unit |
11476288, | Nov 17 2015 | Taiwan Semiconductor Manufacturing Co., Ltd. | Infrared image sensor component manufacturing method |
11476372, | May 13 2020 | Apple Inc. | SPAD-based photon detectors with multi-phase sampling TDCs |
11515014, | Feb 21 2020 | HI LLC | Methods and systems for initiating and conducting a customized computer-enabled brain research study |
11607132, | Mar 20 2020 | HI LLC | Temporal resolution control for temporal point spread function generation in an optical measurement system |
11630310, | Feb 21 2020 | HI LLC | Wearable devices and wearable assemblies with adjustable positioning for use in an optical measurement system |
11645483, | Mar 20 2020 | HI LLC | Phase lock loop circuit based adjustment of a measurement time window in an optical measurement system |
11771362, | Feb 21 2020 | HI LLC | Integrated detector assemblies for a wearable module of an optical measurement system |
11819311, | Mar 20 2020 | HI LLC | Maintaining consistent photodetector sensitivity in an optical measurement system |
11857348, | Mar 20 2020 | HI LLC | Techniques for determining a timing uncertainty of a component of an optical measurement system |
11864867, | Mar 20 2020 | HI LLC | Control circuit for a light source in an optical measurement system by applying voltage with a first polarity to start an emission of a light pulse and applying voltage with a second polarity to stop the emission of the light pulse |
11877825, | Mar 20 2020 | HI LLC | Device enumeration in an optical measurement system |
11883181, | Feb 21 2020 | HI LLC | Multimodal wearable measurement systems and methods |
11903676, | Mar 20 2020 | HI LLC | Photodetector calibration of an optical measurement system |
11903713, | Dec 21 2018 | HI LLC | Biofeedback for awareness and modulation of mental state using a non-invasive brain interface system and method |
9200953, | Jul 28 2011 | STMicroelectronics S.r.l. | Spectrometer including a geiger-mode avalanche photodiode |
9871067, | Nov 17 2015 | Taiwan Semiconductor Manufacturing Co., Ltd. | Infrared image sensor component |
9991417, | Jul 31 2015 | International Business Machines Corporation | Resonant cavity strained III-V photodetector and LED on silicon substrate |
Patent | Priority | Assignee | Title |
4945227, | Nov 25 1986 | The Secretary of State for Defence in Her Britannic Majesty's Government; SECRETARY OF STATE FOR DEFENCE IN HER BRITANNIC MAJESTY S GOVERNMENT OF THE UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND, THE, A BRITISH CORP | Avalanche photodiode quenching circuit |
6384663, | Mar 09 2000 | Politecnico de Milano | Circuit for high precision detection of the time of arrival of photons falling on single photon avalanche diodes |
20030165207, | |||
20040106265, | |||
20050006678, | |||
20050255649, | |||
20060131480, | |||
20080128631, | |||
20090065704, | |||
WO2007127607, | |||
WO2008011617, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Jul 13 2009 | The Regents of the University of California | (assignment on the face of the patent) | / | |||
Sep 15 2009 | FINKELSTEIN, HOD | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 | |
Sep 20 2009 | ZHAO, KAI | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 | |
Sep 30 2009 | ESENER, SADIK C | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 | |
Sep 30 2009 | CHENG, JAMES | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 | |
Sep 30 2009 | YOU, SIFANG | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 | |
Oct 05 2009 | LO, YU-HWA | The Regents of the University of California | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 023359 | /0375 |
Date | Maintenance Fee Events |
Jul 28 2017 | M2551: Payment of Maintenance Fee, 4th Yr, Small Entity. |
Sep 20 2021 | REM: Maintenance Fee Reminder Mailed. |
Mar 07 2022 | EXP: Patent Expired for Failure to Pay Maintenance Fees. |
Date | Maintenance Schedule |
Jan 28 2017 | 4 years fee payment window open |
Jul 28 2017 | 6 months grace period start (w surcharge) |
Jan 28 2018 | patent expiry (for year 4) |
Jan 28 2020 | 2 years to revive unintentionally abandoned end. (for year 4) |
Jan 28 2021 | 8 years fee payment window open |
Jul 28 2021 | 6 months grace period start (w surcharge) |
Jan 28 2022 | patent expiry (for year 8) |
Jan 28 2024 | 2 years to revive unintentionally abandoned end. (for year 8) |
Jan 28 2025 | 12 years fee payment window open |
Jul 28 2025 | 6 months grace period start (w surcharge) |
Jan 28 2026 | patent expiry (for year 12) |
Jan 28 2028 | 2 years to revive unintentionally abandoned end. (for year 12) |